Will artificial intelligence be as transformative to our working world as we might believe? 

Having worked in tech recruitment for 25 years, I’m no stranger to emerging technologies and how they’ll impact the market. AI is the latest that possesses the potential to automate processes, responsibilities and even roles. A lot is possible in theory, but what’s realistic? 

To explore the possibilities and considerations in the world of recruitment and similar areas, I spoke to James Hutt, a consultant at Paradigm Junction, who’s helping organisations navigate AI disruption. 

How will AI change the recruitment process? 

Thanks for joining us again, James. Let me start by describing what some of our software vendors think the future recruitment process might look like. A customer calls a recruitment consultant to discuss an open role. In real time, an AI tool transcribes the conversation and creates a job specification. An algorithm then searches for relevant candidates and alerts them, asking whether they’re interested. So, almost by the end of the call, you’ll have a shortlist. It’s an interesting concept - what are your thoughts? 

I imagine that software vendors would be quite happy to design a world where there was no human touch at all! 

Veterans of digital transformation over the last 10, 20 years would tell you there's a pretty big difference between what's possible in theory, and what would work in reality at companies that already have existing relationships and data. Yes, we're clearly going to see quite a lot of changes to ways of working - new workflows and tools that make a big difference to the experience for recruitment consultant and clients. However, the idea that this is all going to be completely automated is just a dream at the moment. 

Instead, it's interesting to look at where bits of the process can be automated or improved quickly. How are people going to use tools to gain an advantage, rather than simply removing humans from the loop entirely? And how does the market change for everyone else, once some recruiters are using AI tools.  

Is that something that people want? Where are humans happy working with AI technology and where do they still really value that human interaction? 

Let’s look at the example of scheduling interviews. There is research that shows people enjoy picking a time so that they have control, so many candidates would prefer an automated system than a back and forth. However, it’s not technology that will determine whether your company is willing to do that - it's up to you to decide whether to offer up that control. That mutual negotiation to find a time is when you really get to feel each other out. It's a bit like a first date. So, I'd be careful about automating in that space because you're talking about people's first impression of you. 

Similarly, with job specifications, there’ll be tools that can help hiring parties write postings but it's going to be up to humans to add a bit of sparkle and make them stand out. If everyone has access to the same tools, you might think all job ads will look the same. But experience and expertise will be redirected to helping companies stand out against a backdrop of auto-generated job specs. Consultancies will have the expertise in promoting jobs that people want to apply for and ultimately getting the best talent that way. Counterintuitively, that might be when we see people rely on consultancies more, rather than just going to jobs boards. 

I think we'll see some aspects of the opening conversation between consultant and client be automated, such as through transcribing. But how you learn about each other involves connecting with people. 

In terms of getting a holistic view of the candidate beyond the CV, what might that look like over the next five years or so? 

One of the best places to be looking at are Sales Development Rep bots, or SDR bots. These are tools that marketers and people in sales are starting to use to speed up their workflow. Companies often outsourced a lot of that function to research companies and their employees – often overseas in lower cost countries. Then someone who knew the market would look at that summarised information and decide what to do with it. Now the research could be done by these bots. 

I think we'll see the emergence of tools like this in the recruitment industry. People have ever-increasing footprints online. Lots of people in professional roles have LinkedIn profiles and you'll be able to automatically pull in information about companies they've worked at before, as well as similar companies. 

You'll have the opportunity to look much further and wider and to identify what looks like a relevant match for your role. 

We've seen historically how technology can have biases based on the input. What's your view of this in the recruitment process? 

There have been lots of stories about biased outputs of the major AI models, and all the big providers work hard to train these out once attention is drawn to them. All the models are now better at representing the diversity of today’s workforce, at least for overt protected characteristics. There are still challenges about how they interpret names, or subtle signals about someone’s background, as we have seen with implicit biases in humans.  

This comparison with humans, though, is right in a number of ways. We know that humans can be biased and so we create training and systems to reduce and then mitigate these problems when we are making decisions in the workplace. We should do exactly the same with AI tools – evaluate that they have been well trained and ensure that there is oversight and justification for any decisions that they make. One benefit of the technology based systems is at least that they are much more measurable, so we should be able to prove that they aren’t unfairly discriminating, rather than taking it on trust or hope alone.  

Regarding candidate outreach, is there a demographic or are there certain areas where you can utilise this more effectively than others? 

When it comes to outreach to candidates, I think ‘personalisation’ is going to be the big word. You're going to have the ability to send to messages that are way more convincing and that might not even seem automated to some people. 

An example that I came across recently is that OpenAI have partnered with the jobs board Indeed, who are trying to get better responses when they do exactly this. They've built tooling that allows them to look at a candidate's profile and previous experience and personalise those messages. 

So instead of saying, “Hey, are you interested in this job? We think it could suit you.”, they can say, “Hey, this company has told us that they're looking for sales executives who've been working on this type of SaaS platform in the last few months. We've seen that you've been doing this and think you've therefore got a great chance if you apply.” Those types of messages are going to get way higher degrees of engagement and do a way better job of tapping into pools of talent that maybe aren't actively looking for jobs. 

This is going to be a challenge. We've talked before about seeing messages on LinkedIn that feel personal but as soon as you've seen five of them, you recognise that it fits into a pretty common pattern. After that, you start ignoring them. 

Think about the competition that's going to go on here. There is going to be tooling that allows recruitment consultants, but also every other type of outbound marketing under the sun to personalise messages. Eventually, you'll likely develop an intuition for whether it’s automated or whether there's a real sense of humanity behind it. Marketers, including recruiters, will keep iterating and developing new creative ways of getting attention, but the AI bots will likely imitate these quickly. It’s going to be a never-ending battle for attention.  

That's the arms race that we're going to see over the next few years. And I think the answer will ultimately be human relationships, such as a message from someone you've been in touch with for some time. That history can’t be faked.  

I can already recognise that fatigue - I know when I'm getting an AI-generated messages and I automatically ignore it. I'll be interested to see whether some of these technologies will be as transformative as some organisations are claiming, just because humans’ behaviour will change too. 

People often talk about technology being transformative as whether they can imagine the future being really different and, if that doesn't happen, it's not been. I'd say it's a different world where the end state balance can look almost the same. There's roughly the same number of messages, or the same number of job applications. That doesn't mean that the tech isn't having an impact. It just means we reach a new kind of equilibrium. 

However, if you're not using the technology to support you, then you're not even going to be able to play. You won’t get to that level where you can have a human do sensible outreach. It's not that the world will look completely different - it’s that you're going to need the tools to play in the game at all. 

To explore the right solutions for your organisation’s needs regarding AI and other tech solutions, get in touch with Hays or Paradigm Junction today. Read my previous conversation with James on AI adoption here.